Low Comp Accuracy? How to Run Better Comps in Any Market
Your after-repair value is only as good as the comparables that support it. An ARV based on poor comps is not an estimate. It is a guess wrapped in a number that feels precise. If your buyers keep telling you your ARV is wrong, or if your deals consistently fail to sell at your asking price, the problem is almost certainly in how you are running comps.
Running accurate comps is part science and part art. The science is using the right data, the right filters, and the right radius. The art is knowing when to make adjustments, which comps to include, and which to exclude. Here is how to improve both.
The fundamentals: what makes a good comp
A comparable sale (comp) is a recently sold property that is similar enough to your subject property to serve as a reliable indicator of value. The key word is "similar." A comp needs to match your subject on several dimensions to be useful:
- Location: Same neighborhood or subdivision is ideal. Same zip code is acceptable. Different side of town is not a comp.
- Property type: Single-family to single-family, townhouse to townhouse, condo to condo. Do not comp a single-family house against a condo or duplex.
- Size: Within 20% of the subject's square footage. A 1,200 sqft house should be comped against 960-1,440 sqft houses. A 3,000 sqft house is not a comp for a 1,500 sqft house regardless of how close it is.
- Bedroom/bathroom count: Same or within one. A 3/2 comps best against other 3/2s or 3/1.5s. A 2/1 is a fundamentally different product than a 4/3.
- Age: Similar era of construction. A 1960s ranch comps differently than a 2020 new build, even if they are the same size and in the same neighborhood. Build quality, layout, and buyer expectations differ by era.
- Condition: This is the most commonly overlooked factor. A fully renovated comp does not accurately represent the value of a house that needs $40K in work. Match condition to condition, or make explicit adjustments.
- Recency: More recent is better. Comps from the last 3-6 months are strong. Comps from 6-12 months are acceptable. Comps older than 12 months are risky because market conditions may have changed.
Technique 1: Start tight and expand
Begin with the tightest possible search criteria: same subdivision, same bedroom count, same property type, sold within the last 6 months, within 0.25 miles. If this returns 3-5 good comps, you are done. If it returns fewer than 3, gradually expand: increase the radius to 0.5 miles, then 1 mile. Extend the date range to 9 months, then 12 months. Broaden the size range.
The reason to start tight is that the closest, most recent, most similar sales are the most accurate indicators of value. Every expansion you make introduces more noise into your estimate. A comp from the same street that sold last month is worth more than five comps from a mile away that sold nine months ago.
Technique 2: Use sold prices, never list prices
This seems obvious but it trips up beginners regularly. The list price is what the seller hopes to get. The sold price is what the market actually paid. In many markets, there is a significant gap between the two. A property listed at $210K that sold for $185K tells you the market value is $185K, not $210K. Some investors mistakenly use active listings or pending sales as comps. Active listings are useful for understanding the competitive landscape but they are not evidence of value because the transaction has not closed.
Pending sales are closer to useful, but the final sold price is often lower than the pending price due to inspection negotiations, appraisal adjustments, and closing credits. Use closed sales as your primary data source and active listings only as supplementary context.
Technique 3: Adjust for condition differences
The biggest source of comp inaccuracy is failing to account for condition. If your subject property needs a full renovation and your best comps are all fully renovated, those comps tell you the after-repair value but not the as-is value. Conversely, if your comps include distressed sales (foreclosures, short sales, investor flips purchased below market), those will understate the value of a renovated property.
When adjusting for condition, think in categories:
- Fully renovated: These comps support your ARV directly (what the property will be worth after repairs)
- Good condition (needs minor updates): Subtract $5,000-$15,000 from these comps to approximate full renovation value, or add that amount if your subject is in similar condition
- Fair condition (needs moderate work): These require larger adjustments and are less reliable as ARV indicators
- Distressed/as-is: These comps support your as-is value but are poor ARV indicators. They are useful for determining purchase price, not sale price after rehab
The best comp set for an ARV estimate includes at least 2-3 fully renovated or recently remodeled sales. If you cannot find renovated comps, your ARV estimate will have a wider margin of error, and you should price more conservatively.
Technique 4: Account for market trends
Real estate markets move. In an appreciating market, comps from 9 months ago understate current value. In a declining market, they overstate it. If you are using comps from more than 6 months ago, check the direction of the market by comparing the average price per square foot of recent sales (last 3 months) against older sales (6-12 months ago).
If average price per square foot has increased 5% over the last 6 months, a comp from 9 months ago that sold for $150/sqft might support $157/sqft today. If it has decreased 5%, a comp from 9 months ago at $150/sqft might only support $142/sqft today. These adjustments matter, especially on higher-priced properties where a 5% difference is tens of thousands of dollars.
Technique 5: Use multiple data sources
No single data source captures every sale. The MLS captures sales that went through real estate agents, which accounts for the majority of transactions. But off-market sales, for-sale-by-owner transactions, and investor-to-investor deals often do not appear in the MLS. County deed records capture all property transfers regardless of how they were sold, making them a useful supplement.
Cross-referencing MLS data with public records data helps you catch sales that one source might miss and verify that the data in both sources agrees. If the MLS shows a sale at $185K and the deed records show $185K, you have high confidence. If there is a discrepancy, investigate. Sometimes the MLS sold price includes seller concessions that are not reflected in the deed, or the deed records have not been updated yet.
Technique 6: Comp by price per square foot, not total price
Total price comparisons are misleading when properties differ in size. A 1,200 sqft house that sold for $180K ($150/sqft) and a 1,800 sqft house that sold for $225K ($125/sqft) are both in the same neighborhood, but using the $225K sale as a comp for your 1,200 sqft subject would overstate its value.
Price per square foot normalizes for size differences and gives you a more reliable basis for comparison. Calculate the price per square foot for each comp, take the average or median, and multiply by your subject property's square footage to estimate value. This is not a perfect method (homes do not scale linearly, as a 3,000 sqft home does not sell for exactly twice a 1,500 sqft home), but it is more accurate than raw price comparisons.
Technique 7: Know what to exclude
Not every sold property is a valid comp. Exclude the following from your analysis:
- Family transfers: Sales between family members (identifiable by same last names on buyer and seller, or sales at $1 or $10) do not reflect market value
- Foreclosure and short sales: These typically sell below market value due to the circumstances of the sale. Use them for as-is valuation but not for ARV
- Flips that just sold: A property that was purchased by an investor 4 months ago for $100K and just sold for $200K after renovation might overstate ARV if the renovation was unusually high-end for the neighborhood. Look at what finishes were used
- Outliers: Any sale that is dramatically above or below the cluster of other comps. A sale at $250K in a neighborhood where everything else sells for $160-$180K probably had unique circumstances (large lot, custom addition, buyer overpaid)
- Different school districts: School district boundaries can create sharp value differences within short distances. Two houses 500 feet apart in different school districts can have 10-20% value differences
Technique 8: Verify with active listings
After you have your comp-based ARV estimate, cross-check it against currently active listings in the area. If renovated properties similar to yours are listed at $195-$210K and your ARV estimate is $215K, you may be too high. Listings that have been active for more than 30 days without going under contract are especially telling. They indicate a price the market is not willing to pay.
Conversely, if similar properties are going under contract within a week of listing, the market is absorbing inventory quickly, which supports the higher end of your comp range.
When comps are scarce
Some markets and property types simply do not have enough recent sales to build a reliable comp set. Rural areas, luxury properties, and unusual property types (converted churches, large acreage parcels, mixed-use buildings) all present this challenge.
When comps are scarce, expand your methodology:
- Extend the radius but stay within the same market character (do not cross from urban to suburban)
- Extend the time frame to 18 months but apply a market trend adjustment
- Use pending sales (with a 3-5% discount) as supplementary data points
- Talk to local agents who specialize in the area, as they have context that data alone cannot provide
- Be conservative in your estimate and build in a wider margin of error
In markets with scarce comps, communicate the uncertainty to your buyers. Presenting your ARV as "$175K-$190K based on limited comparable data" is more honest and more credible than a single number presented with false confidence.
The accuracy standard
Professional appraisers target ARV estimates within 3-5% of actual sale price. As a wholesaler, your goal should be within 5-10%. On a $200K property, that means your ARV should be within $10K-$20K of what the property would actually sell for after renovation. Consistent accuracy in this range builds trust with your buyers and ensures your deals are priced correctly from the start.